PDA

View Full Version : Athlon XP Unlocking...


pelly
09-22-02, 09:13 PM
What materials/kit do you prefer to use to unlock your Athlon XP?

Conductive pen? Defogging kit? A specialized kit? Peanut-Butter?

You get the picture....
I'm snagging a new CPU and would like to know the best way to unlock the bridges...

Thanks ahead of time guys!

:D

thcdru2k
09-22-02, 09:21 PM
ehh i have an athlon tbird 1 ghz and a xp 1.4ghz that i both forgot to unlock before i ac3 the heatsink and fan to it and stuck it into a motherboard. :( oh well i'm a dumbass. i just get so excited when they come in the mail, i just stick them in half hazardly.

volt
09-22-02, 09:34 PM
Originally posted by pelly
What materials/kit do you prefer to use to unlock your Athlon XP?

Conductive pen? Defogging kit? A specialized kit? Peanut-Butter?

You get the picture....
I'm snagging a new CPU and would like to know the best way to unlock the bridges...

Thanks ahead of time guys!

:D

I'm with you brother!
I haven't unlocked my Athlon either heh

koslov
09-23-02, 02:44 AM
I've tried using that rear window defogger repair paint, but I ended up getting frustrated and quitting.

The crappy paint stuck to the tape instead of the ceramic and pulled right off when I tried to mask another bridge. That and it flows to easily, going everywhere where you don't want it to.

I am planning on trying again with conductive ink, but I am still smarting from wasted time and money from the first attempt.

fasedww
09-23-02, 06:11 AM
[QUOTE]Originally posted by koslov
I've tried using that rear window defogger repair paint, but I ended up getting frustrated and quitting.

The crappy paint stuck to the tape instead of the ceramic and pulled right off when I tried to mask another bridge. That and it flows to easily, going everywhere where you don't want it to.

I am planning on trying again with conductive ink, but I am still smarting from wasted time and money from the first attempt. [/

Go to HighSpeedPC.com and get there XP Ulocking kit, The conductive grease is easy to apply with a pin needle and magnifyed Glass.:D

pelly
09-23-02, 08:10 AM
Thanks guys...keep those suggestions coming....( I'm off to go take a look at that kit... )

:D

john19055
09-23-02, 08:39 PM
I used the rear window defogger bown stuff and clear tape and a big magnifiing glass and a ohm meter and it was a pain in the asss but after a couple of hours with a steady hand when I put it back in everthing works great.

Switch
09-24-02, 07:08 PM
hmmm... Unlocking was really fun for me. At first I was trying to do a "easily removed" job but then after hours I said screw it. My first try was to make my own trace tape. I turned a peice of tape upside down and applied defogger repair stuff in rows to the sticky side. Then just stuck that onto the processor. Still wanting to use the clear tape I cut a VERY thin peice of about half an inch in length and used it to cover the pits. Once the tape was secure I just painted right over the bridges with the defogger repair. No messy goo or glue to fill the pits with. If I hadn't of attempted the trace tape idea first, the whole operation would have taken less than 3 minutes.

SnakeEyes
09-25-02, 10:27 AM
I've got the 2100+, so unlocking it was a bit more of an experience for me than with 2000+ and lower. With those, it's mainly a matter of connecting the LV1 bridges. With the 2100+. since most overclockers want to increase the FSB, lower multiplers are desirable. But simply connecting the LV1's doesn't get multipliers lower than 12.5. It actually required a dremel engraver, a multimeter, some of the rear-window defogger, super-glue, a hobby knife, and some magic tape.

The actual unlocking process only ended up taking me about 30 minutes, though I had to:
1) Use the engraver on it's lowest setting to break a bridge (LV3 I think).
2) Test the points on the bridge I was attempting to break with the multimeter, checking for continuity.
3) Repeated steps 1 and 2 three times, until the multimeter registered no continuity. (It was important to etch very shallowly on each pass, as I only wanted the bridge connections broken, and there ARE other layers below them in the die. Break those, and the processor's most likely toast).
2) Use the magic tape (Scotch brand), two 1" long pieces, to isolate the laser-etched groove between the points of each set of bridges.
3) Apply superglue (carefully) to the gap between the tape, in the laser-etched groove, effectively sealing and insulating the grounds at the bottom of the pits.
4) Use the hobby knife to scrape any excess superglue from the surface of the die, including any that gets on the LV1 points or the ground dots nearby.
5) Use two more strips of magic tape to isolate a pair of points in the bridges to be connected.
6) Used a fine-point toothpick to apply a thin amount of the rear-window-defogger-repair goop to the gap between the strips of tape, connecting the two points of that part of the bridge.
7) Let the goop set for 1 - 2 minutes (no longer, or it pulls off with the tape, and no shorter, or it's too wet and runs, possibly connecting with another bridge or one of the ground points next to them on the die).
8) Remove the tape.
9) Allow an extra minute or two for the goop to set harder, so that it isn't messed up by the tape that is going to be on top of it.
10) Repeat steps 5 - 9 until all points are connected on the LV1 bridge, as well as on the proper pair from the LV3 (enablind lower multipliers).
11) When all of the goops dried and in place on the bridges, I cut two small, narrow pieces of tape the same size as the modified bridges, and applied it over them, to protect the goop from scrapes and scratches that could break the connections.
12) Insert processor in motherboard, apply heatsink compound (Artic Silver II, in my case), attach heatsink, put the rest of the system back together.
13) Start playing with multiplier / fsb settings in bios, attempting to find the best overall combination for performance with the processor. This step took longest (2 weeks of playing.. several reformat/reinstalls of WinXP, thanks to the hdd controller not liking certain bus speeds and corrupting the data.., etc.)

Since I did my unlock, I've read of several alternative methods. One involved using some sort of conductive rubber (has parallel rows of conductive wire in the rubber) cut into a thin layer and placed on the LV1 bridges, (held with tape) to make the contacts. Pin tricks on the bottom of the processor (er, maybe for voltages, not sure, lol). Etc.

There are many methods, but as far as I know, the most common involves sealing the laser-pits with something, then connecting the bridges with some other conductive goop. BTW, pencil lead DOES NOT work, nor does trace tape, at least with my 2100+. Neither conducts well enough (too much resistance).

In the future, I hope AMD does what was done with my 1.33 OEM processor- ships them unlocked from the factory. Sure as heck beats the sweaty job of cutting traces and applying foreign substances to the top of the die. (sweaty from nervousness at the possibility of ruining the processor :D)

pelly
09-25-02, 12:21 PM
Well.....I think I've done it....

I purchased a de-fogger kit from PepBoys and went to town on my nice new green XP1600...

First, I taped-off the contacts for the L1 bridges. Stealing a bright orange crayon from Heather ( my fiance ), I began coloring over the gaps or pits between the contacts. Soon, the gaps were filled as I could see the orange wax building up.

From here, I tried "masking" the trace-lines and using the de-fogging kit....Unfortunately, the damn Scotch-tape kept pulling up my beautiful new traces when I moved on to the next contacts.

As a result, I decided to pull a McGuyver move...I took the fat little applicator brush that came in the kit and dipped it into the bottle of conductive "paint" and left it out in the open to harden. Once the brush was as hard as a brick, I cut a knife-shape into the head of the applicator brush using a pair of scissors. From this point on, I could "draw" the traces with ease and didn't need to use tape anymore...

I'm not sure if it works yet, though I'll be assembling the new system once I get home tonight....

I'll let you know how it goes....thanks for all your help guys!!!

:D

SnakeEyes
09-25-02, 01:17 PM
Cool trick, Pelly! I'll keep that in mind next time I unlock a processor, since I spent a LOT of time just getting the masked areas right so that I wouldn't have my blob of gunk getting where I didn't want it. BTW, The Pep Boys' defogger kit was the same one I bought. :D

pelly
09-25-02, 08:06 PM
Tee hee hee.....it worked on my first try!!!
( Sigh of relief )

So far, I've had this XP1600 up to 1730MHz...I'll play around more later tonight....

Life is good.....I deserve a beer....

:D

SnakeEyes
09-25-02, 10:58 PM
Hehe. That's just past 2100+ speed Pelly. (1729).

My 2100+ is only good for a (truly) stable 1845. Even though it seems stable at 1870, running Prime95 says differently (rounding errors). I've even managed to boot into WinXP at 2000+Mhz one time. I'm not sure how it worked, since I can't reproduce that. That's also similar to my one and only successful 3DMark2K1SE completion at 1904MHz (can't get it to stay stable long enough to do that again now either, but the results are out on MadOnion to prove I did it one time. :p)

Keep the reports coming! (BTW, what you've hit already is certainly no small acheivement, but I'd guess you're gonna end up close to where I am at 1800+ MHz).

Smokey
09-26-02, 05:11 AM
Originally posted by pelly
Tee hee hee.....it worked on my first try!!!
( Sigh of relief )

So far, I've had this XP1600 up to 1730MHz...I'll play around more later tonight....

Life is good.....I deserve a beer....

:D

Thats a super overclock there :D Can you let us know what your settings are, like FSB, cpu and also what cooling you are using? The reason, I still havent unlocked my 1800+, I tried but I dont have a magnefying glass in the house, so it was a bit hard and didnt work :( Reading things like this though only makes me want to try harder :p

pelly
09-26-02, 08:33 AM
I'm at work right now, so I don't have access to my piece of paper with all the calculations on it ( multiplier/FSB as well as CPU stepping ).......but I'll have a mini-article up soon....part of a larger project...

The goods being used are:

XP1600+ w/ green PCB
MSI KT4 Ultra
(2) 256MB Corsair XMS PC3200 CAS2
OCZ Goliath 2 SE w/80mm Fan
Tweakmonster Shim


:D

pelly
09-28-02, 09:18 PM
Here's the core information....

http://pubpages.unh.edu/~seanp/PC/CPU_Core.jpg

11x multiplier...158FSB... --> 1738

:D

Once I get some more time, I'll try for the insane FSB speeds.....This was just aquick test though with good results....

:)

pelly
09-28-02, 09:19 PM
Keep your eyes peeled for a related article showing up in the next few days...

:D